Adding applications

About this task

You have to add a new application to start using it. Only cluster and consumer administrators, or users with Application Management permission can add an application.

When you add a MapReduce application, you are registering with the system a simple application that includes some default values, such as those for resource groups and the resource plan policy.

Typically, you add an application when:
  • The consumer does not exist. When an application is added, the system creates a leaf consumer with a name identical to your application name under the MapReduceConsumer node.
  • The consumer exists but does not have any applications registered to it.
  • The consumer exists but applications registered to it are disabled.
  • The consumer exists but not under the MapReduceConsumer parent node.

Before adding an application, ensure that user accounts have been set up for users who are allowed to use the MapReduce application.

To add an application, follow these steps:

Procedure

  1. Click Workload > MapReduce > Application Profiles.
  2. Click Add.

    The Add Application window displays.

  3. Enter parameters for the new application.
    1. Enter a name for the application. The name you specify must be unique in your cluster.
    2. Specify the default job priority for all jobs submitted for this application. Enter any value between 1 and 10,000.
    3. Select the resources groups for this application. By default, the application is added to the built-in ComputeHosts and ManagementHosts resource groups.
    4. Optionally, add any environment variables for your application; for example, PMR_HOME and HADOOP_HOME.
  4. Click Add.

    The newly created application displays in the application list in the enabled state. A consumer with the same name will now be available under the MapReduceConsumer node at Resources > Resource Groups > Consumers.

    Note some of the default properties for the application you added:

    Property Description Default value
    Policy Specifies the workload scheduling policy used:
    • R_PriorityScheduling for priority scheduling.
    • R_Proportion for proportional scheduling.
    		 R_Proportion, which means that the number of resources is proportional to priority
    Prestart Specifies whether the Application Manager, service instance manager, and service instances are prestarted for enabled applications when MapReduce is started.

    To optimize performance for your MapReduce workload, we recommend that you prestart services.

    		 True
    Scheduling Affinity Determines whether resource-aware or data-aware preference expressions or rack-aware scheduling will be evaluated for jobs and tasks of this application. None
    Standby Service Specifies whether resources that have been allocated to the application will have standby services running when slots are released by the Application Manager. False
    Preemption Scope Determines whether jobs in the application can preempt jobs of lesser or equal rank, or only preempt jobs of lesser rank when preemptive is set to true. Lesser or equal ranked jobs
    Enable Selective Reclaim Selective reclaim feature: "true" if enabled False
    Configured Slots for Preloaded Service Number of preloaded service slots configured for the application. 10000
    Slots threshold for standby services Maximum number of slots that standby services can run on for the application. -1

What to do next

Now that you have added an application, you can update its properties by modifying the application profile.